An HSR Country is a Centralized Country

1950s’ Japan was a fairly monocentric country, in which everything was in Tokyo. When it built the Shinkansen, the expectation was that fast travel nationwide would make it easier to do business in the other cities, reducing centralization. Instead, the opposite happened: the Shinkansen made it easier to get to and from Tokyo, increasing centralization. At the same time, the US, which forwent its rail system and built the Interstate system, saw its manufacturing belt disintegrate, with production moving to the South. If we think of high-speed rail as a nationwide version of rapid transit, then we get the same pattern seen in cities, in which transit works hand-in-hand with centralization.

As with transit, there are exceptions – namely, Germany. Germany’s point-to-point HSR network coexists with its polycentric layout. What this suggests is that HSR does not create centralization so much as reinforces it when it already exists. The Shinkansen made the rest of Japan more dependent on Tokyo, and the TGV has made most of France more dependent on Paris, but that’s because the existing traffic patterns were such that only lines connecting to or near the capital would be competitive. Thus, only connections from a provincial city to the capital are fast, and the loss of province-to-province connectivity (more precisely the deemphasizing of such connections, since both Japan and France continued to build nationwide freeway networks) leads to a loss of independence in the provinces. Lille has redeveloped with the help of the TGV, but this has involved marketing itself as a city close to Paris, Brussels, and London. It’s not the same as Paris itself, which gets by without needing to tout how close it is to London or Lyon.

This does not detract from the fact that HSR can lead to development. Lille really did redevelop with the help of the TGV. Many cities right outside Ile-de-France, such as Tours, are seeing a property boom fueled by fast train links to Paris. The Shinkansen helped bind the Tokaido-Sanyo megaregion, redeveloping cities at appropriate commute range, for example Mishima. The issue here is that a city bound by a megaregion is no longer an independent region, for all that entails.

The consequence in the US is of course not that HSR will turn the country into a single-city country. The US is too big and decentralized. But within each region, HSR is going to bind megaregions together in a way that leads to the same loss of independence. In the Northeast, we can expect the region to be far more dependent on its four primary cities, especially New York. Providence would benefit from being about 20 minutes from Boston and 1:15 from New York, but it would be drawn fully into those two cities’ orbits. The economic development it can expect is not the sort that still clusters in New York and Boston, but rather the lower-end development that is worthwhile to outsource to lower-cost regions. It would be competing with Middlesex County, New Jersey for jobs, rather than with Midtown or even with Downtown Brooklyn. Likewise, in California, we can expect to see more dependence on Los Angeles ans San Francisco, with Bakersfield and Fresno relegated to secondary status.

What I’m doing here is describing in grimmer terms what is cheerfully described as development in various pro-HSR brochures. An advanced economic system, including fast transportation, will lead to specialization, and this includes specialization into center and hinterland. This is new economic geography: reduced transportation costs lead to more rather than less specialization, and HSR reduces transportation costs with respect to time for certain kinds of work.

Ironically, what this implies is that the best way to preserve independence is to not build any binding infrastructure, or engage in national planning. Toronto will remain independent of New York so long as there are separate currencies, separate national markets, and different infrastructure clusters; there’s not much demand for New York-Toronto travel (the air market has about 100,000 monthly passengers in both directions; the top intranational market, New York-Miami, has about a million), so there will not be any new infrastructure between New York and Toronto anytime soon, which will further reinforce those cities’ distinct economies. Montreal, which occasionally seeks HSR to New York as economic development, is doing so explicitly to have an economic basin separate from Toronto’s; it is willing to sacrifice economic independence to achieve some independence from Toronto.

This seems to have been Jane Jacobs’ view in The Economy of Cities and Cities and the Wealth of Nations. Although she wrote about the economic links of the original manufacturing belt megaregion, she wrote even more about the economic links within each city region, and had a dim view of megaprojects; in Cities and the Wealth of Nations, she also rejected national currencies, and proposed city-states as a replacement for nation-states. I have little doubt that she would oppose HSR, just as later in life she came to oppose rapid transit and support jitneys.

Not believing that everything Jacobs said is gospel, I take a more neutral view. The HSR-bound megaregion is more efficient in a way than having ten independent cities along the Northeast Corridor, just as New York is better off today as a single city than it would have been as separate cities if the 1898 amalgamation had not gone through, despite the loss of independence Brooklyn has endured. However, this efficiency is achieved via a brutal division of labor between the cities: some become core, some become periphery.

Of course, this may be sufficient consolation in the small cities that would love to become suburbs of successful cities. At the time of this writing, Fresno’s unemployment rate is 15%, and Bakersfield’s is 14%. The Central Valley is seeking prisons as a form of job creation. Providence is better off, but despite recent economic growth and slow absorption into Greater Boston, it’s one of the higher-unemployment regions in the Northeast. Loss of independence is not necessarily bad. But conversely, the fact that this development is good does not mean that it will really turn the smaller cities into productive city regions; it will just make them more comfortable peripheries of cities in which there’s so much that the residents don’t have to care about intercity travel.


  1. Tom West

    Comparing you comments on France and Japan vs. Germany and USA… what you seem to be saying is that HSR gets built to reflect the biggest travel patterns, and then re-enforces them.

    • Alon Levy

      Yes, exactly. And this comes from the fact that, like rapid transit, HSR has high construction costs, high capacity, and low operating costs, and runs along corridors. So in Switzerland, the major preexisting markets were point-to-point since there were multiple dominant cities, so the growth of intercity rail preserved this polycentricity. In France, the major markets were province-to-Paris, and the TGV has cemented Paris’s dominant role. It’s not like airports, which are far more flexible, at the cost of much higher operating costs at short distances and much lower capacity.

  2. Loren Petrich

    I’ve also considered this question, and I’ve found several network topologies for HSR networks of countries with more than 1 or 2 major lines or served cities.

    France and Spain both have hub-and-spoke systems centered on their capitals, Paris and Madrid, and some of France’s HSR lines cross borders into the UK, Belgium, and Holland. I remember coming across someone’s claim that Spain’s HSR is partially political, to help unify Spain.

    Japan and Italy have linear networks with branches, because of the shapes of their nations. Japan’s network extends from Kagoshima to Aomori, the entire length of three of Japan’s main islands, and it will be extended to Hakodate and eventually to Sapporo, in the northernmost main island. Italy’s network is a flattened S between Turin and Naples, and eventually to Bari. Better-developed UK HSR would also have that topology, extending from London to Birmingham, Manchester and eventually to Edinburgh and/or Glasgow.

    China and Germany have irregular multiply-connected networks, and rather patchy ones at that. However, China is likely to fill in much of the patchiness over this decade. Germany’s construction pace is, however, slower.

    • Alon Levy

      In the US, the network would not be simply-connected, either. But it would be more like a bunch of (usually) simply-connected regional networks, glued together at the edges. So the Northeast would have the NEC, Empire, and Keystone (and maybe Boston-Albany, breaking simple connectivity), and the Midwest would have the Chicago hub network, but the two networks would meet up in Pittsburgh and Buffalo to provide good service on intermediate markets like Philadelphia-Cleveland. The results would also be strongly regional: the Northeast would grow more dependent on the four core cities and especially New York, the Midwest would grow more dependent on Chicago, the Southeast would grow more dependent on Atlanta, Texas and the nearby areas would grow more dependent on Dallas and Houston. Cities at the boundary would either be drawn into one region (which is what I think would happen to Buffalo and the cities of Virginia) or be able to grow as in-between cities dependent on more than one core, as Lille has (which is what I think would happen to Pittsburgh).

      Tellingly, although the Japanese Shinkansen network is a straight line with two underperforming branch lines, it’s fostered dependence on Tokyo nonetheless. It’s not like in France, where the TGV network is such that I’ve heard the phrase “French train metric” used in a math lecture to describe a star-shaped metric, much like how the Manhattan metric is the distance along a grid. Tokyo’s distinguished position is economic rather than geometric.

      • anonymouse

        Note that the Japanese HSR network is not actually a straight line. It’s two straight lines, both heading straight to Tokyo. There is no actual connection between the two systems, though they share a station and are next to each other. So you can’t even in theory have a Shinkansen train run through Tokyo, even if the different corporate ownership and power supplies were not a problem.

        • orulz

          The power supply is not a problem since rolling stock already exists that can transition from 50 to 60 hz (the Nagano Shinkansen, currently being extended into the Hokuriku Shinkansen) has parts of the line that run at 50 and others that run at 60.

          A physical connection would be trivially easy to construct between tracks 14/15 of the Tokaido Shinkansen and 22/23 of the Tohoku Shinkansen.

          The only barrier to doing so is the different corporate ownership. And given the dominance of Tokyo, most through traffic would only be using the link to get from the Tohoku Shinkansen to Shinagawa and perhaps Shin Yokohama, or from the Tokaido Line to Ueno. I suspect the market for train travel between Aomori and Osaka, for example, is small.

      • Anon256

        I’ve more often heard that called the British Rail metric (even though the UK network is less centralised than SNCF) or the Napoleon metric (since the centralisation predates the railways).

        Tokyo’s distinguished position is cemented by the total lack of through-running on Shinkansen trains. (I think it isn’t even physically possible for trains to run through between Tokaido and Tohoku?) Given the high level of pre-existing centralisation there may not be much demand for it, but it’s still surprising they don’t run through as far as Shin-Yokohama/Omiya at least for the same reasons you argue for commuter through-running at Penn Station.

        • Alon Levy

          Well, their commuter trains run through from Yokohama to Omiya, along two different lines (Keihin-Tohoku, and Shonan-Shinjuku). Between Shinagawa and Ueno, there’s also the Yamanote Line.

          If I’m not mistaken, there were plans to run Shinkansen trains through in the JNR days, but by the time the Tohoku Shinkansen reached Tokyo, JNR had already been split, and the extra tracks that were supposed to be used for through-service were used by JR Central in order to increase turnback capacity.

          • Anon256

            Differing corporate ownership doesn’t seem to interfere with through running on urban Japanese lines, or even on the Shinkansen between JR Central and JR West, and between JR West and JR Kyushu. And if the arguments about Penn Station generalise, through-running would increase capacity more than using the same number of tracks for turnback.

            In Germany the ICE trains generally run through to Gesundbrunnen/Ostbahnhof/Sudkreuz (whichever is on the far side) in Berlin, to Altona in Hamburg, etc. This seems like a good system and I’m surprised the Japanese don’t do it in Tokyo.

          • Alon Levy

            Yes, the capacity would be increased, you’re right. The new Tokaido Shinkansen platforms may not be well-aligned for it though, I’m not sure.

            In either case, I suspect the main barrier is agency turf. JR Central wants all trains using the Tokaido Shinkansen to have exactly 1,323 seats, in order to allow maximum flexibility in case rolling stock has to be replaced and also maximize capacity. JR East’s single-deck trains have fewer seats, since they are not single 16-car sets but rather 10-car sets coupled to 6-car sets. It’s also very hard to have a station in the middle of a track pair that’s used both as a through-station and as a terminal in both directions. If it’s a terminal in just one direction, then the terminating tracks go in the middle and the through-tracks go around them, or if it’s two independent track pairs (e.g. Penn Station) then it’s no problem, but terminal from both directions plus through-service is hard. Best that can be done is schedule some JR East trains to Shin-Yokohama using the extra capacity gained by the lower speed and then terminate on new tracks (expensive), and do the same with JR Central at Omiya (also expensive).

          • anonymouse

            Looking at Google Maps, on the north side, at least one, if not both of the two westernmost Tohoku tracks can be connected through to the station throat of the Tokaido side, and on the south end, pretty much all of the Tokaido tracks can be connected through, but there’s some kind small yard or something down there that would have to be moved out of the way. As far as the power supply goes, I suspect that they’d have to install switchable power supplies on the station tracks, as there might be a possibility of a long train with an internal power bus bridging the two systems. I suspect that the reason for the lack of connection also has to do with the fact that there are plenty of non-connected private railway lines all over Japan, including ones with incompatible track gauge.

            As far as Berlin goes, the reason for trains running through is quite simple: there’s no layover space at Hauptbahnhof. Especially on the upper level with only four tracks, but even on the lower level’s eight tracks, capacity must be a bit tight and it’s easier to just take the train to the nearest place where it can be serviced, which are just outside the S-Bahn ring, so you may as well let passengers stay on until the transfer station at the ring.

          • Alon Levy

            I sincerely think it boils down to different operating patterns. The Tokaido Shinkansen is at capacity, so JR Central wants to run single full-length sets on it with the same number of seats, for easy swapping. The Tohoku Shinkansen is nearing capacity south of Omiya but still has some space, and branches to the north so that different train lengths are desired, including the coupling and uncoupling moves for the Mini-Shinkansen.

  3. Loren Petrich

    Looking at North American HSR proposals, they fit all these topologies.

    The eastern half of North America south of southern Canada could be unified as one big HSR network, but it would have a topology much like Germany’s and China’s, and it would be very patchy, just like theirs. The western half is very thinly populated and very mountainous, making it bad territory for HSR. Though the west coast is also mountainous, but it has long linear valleys, and it is very populous.

    Only the Chicago area would have a hub-and-spoke topology; all the others would have linear or branched-linear topologies. The Northeast Corridor is branched-linear, and a Southeast-Corridor extension into North Carolina and Georgia would continue this approximately-linear topology. Eastern Canada’s proposed system, California’s planned system, and the Pacific Northwest’s emerging system are all linear or branched-linear.

    Concerns about rail-induced centralization are far from new; there is a famous NIMBY poster from 1839 Philadelphia that warned that that city would become “a suburb of New York”:

  4. JJJ

    You cant compare Providence with Fresno and Bakersfield. I feel that with this example, your geography may have failed your analysis.

    Providence is 50 miles from Boston.

    San Pedro (within the city of LA) to Van Nuys (also the city of LA) is….40 miles.

    So yeah, it’s no wonder Providence is drawn into Boston…if they were in California, they’d be the same damn city.

    Fresno to LA, meanwhile, is 220 miles. Almost the exact same distance as Boston to NYC, which obviously manage to be very independent of each other.

    In your example, you’d want to substitute in palmdale, which is 60 miles from downtown LA and dreams of being a suburb.

    With these huge distances come many factors that allow the cities to compete based on their geographic advantages. Fresno obviously offers much cheaper housing than LA or SF. $90,000 will buy you a 3 bedroom in Fresno. In SF, a small closet. Thats one huge reason companies might chose to relocate, if the transportation is there to support their needs. And unlike Palmdale, they’d be in a completely independent city, with its own politics and such.

    So yes, HSR will reinforce existing patterns. In the northeast, that means the coast cities, with places like Hartford and Pittsburgh being left behind because theyre not on the spine.

    In California, that spine will be the 99, as it is today.

  5. Benjamin Hemric

    Alon Levy wrote (in part) [editing, added numbering and added emphasis is mine — BH]:

    [1] Ironically, what this implies is that the best way to preserve independence is to not build any binding infrastructure, or engage in national planning.

    [2] Montreal, which occasionally seeks HSR to New York as economic development, is doing so explicitly to have an economic basin separate from Toronto’s; it is willing to sacrifice economic independence to achieve some independence from Toronto.

    THIS [the above points?] seems to have been Jane Jacobs’ view in “The Economy of Cities and Cities” and “the Wealth of Nations.” Although she wrote about the economic links of the original manufacturing belt megaregion, she wrote even more about the economic links within each city region, and [3] had a dim view of megaprojects; in “Cities and the Wealth of Nations,” she also rejected national currencies, and proposed city-states as a replacement for nation-states. I have little doubt that she [4] would oppose HSR, just as later in life [5] she came to oppose rapid transit and support jitneys.
    [6] The HSR-bound megaregion is more EFFICIENT in a way than having ten independent cities along the Northeast Corridor, [7] just as New York is better off today as a single city than it would have been as separate cities if the 1898 amalgamation had not gone through, despite the loss of independence Brooklyn has endured.
    – – – –
    Alon, interesting post and interesting point of view! I hope I’ve understood and summarized (via editing) your post correctly.

    My comments are tentative, as I haven’t re-read the books you mention with your comments in mind. Plus, as mentioned previously, I won’t have time for an extended discussion right now – but do want to put forth a “place saver” for an alternative view point.” (Also let me point out, here, that I think, the “The Question of Separatism,” the “Nature of Economies” and “Systems of Survival” (indirectly) are very helpful in understanding what Jacobs really meant in “Economy of Cities” and “Cities and the Wealth of Nations.”)

    Off hand, it seems to me that the thoughts you are attributing to Jacobs are very different from what she was actually “saying” in these books. In the “Economy of Cities” Jacobs was interested in understanding why and how city economies grow and thrive and stagnate and decay (just as in “Death and Life . . . “ she was interested in understanding why and how cities grow and thrive and stagnate and decay). Likewise, in “Cities and the Wealth of Nations” she was trying to understand, essentially, why and how national economies grow and thrive and stagnate and decay. Here she came to the conclusion, more or less, that it’s not fruitful to focus on “national” economies as the real unit of organization, rather it’s more useful to focus on cities and their regions are, and we do ourselves great harm by ignoring this. (An example of this is having national currencies – or, worse still, international currencies. Although the Euro hadn’t been created yet, I believe, Jacobs was very prescient about why such international currencies don’t work.)

    With regard to 1:

    It seems to me that Jacobs is NOT against what you call binding infrastructure. I believe, in fact, in one of her books (or interviews) she discusses how small political units do not necessarily impede the construction of such infrastructure (with the implication that such infrastructure is good and important). Also, let me say here, that while Jacobs pointed out the problems of over specialization, it doesn’t seem to me that in her view over specialization was due to what you are calling binding infrastructure and its supposed creation of too much interdependence. (Furthermore, in a slightly different context, Jacobs discussed in “Death and Life . . . “ the importance of what you are calling “binding infrastructure” for a system’s (i.e., city’s) health.)
    With regard to 2:
    As you may know, Jacobs actually wrote a book about Canadian separatism, “The Question of Separatism,” where she discusses Montreal quite a bit, naturally. (I only read it about three years ago, when I purchased an amazingly dirt-cheap, mint condition copy from Amazon. The postage cost more than the book! Amusingly, the copy I bought was originally from some state agency of Alaska – and this was when Sarah Palin was running for VP!)
    In Jacobs’ books, she’s not interested in “independence” per se. Rather, the point of independence, basically, is that it provides for better feedback and better accountability – and thus healthier “systems.” So, whether, Montreal has binding links to Toronto, New York, Chicago, etc. is not really relevant.
    With regard to 3:
    I don’t think it’s really true that Jacobs takes a dim view of “mega projects.” Again, remember, she’s interested in understanding why “systems” grow and thrive or stagnate or decay. Some mega projects are helpful, others aren’t. I don’t believe Jacobs has ever said, for instance, that mega projects like the Suez canal, or the Panama canal – or the NYC subway system, either – were “bad” projects.

    With regard to 4:
    I don’t know how Jacobs felt about high speed rail (HSR), but it doesn’t seem to me that her concerns are similar to ones you are presenting – i.e., whether they, as you say, “economically” bind cities together and, supposedly, make cities less “independent.” It should be remembered that in “Death and Life . . .” Jacobs actually points out the economic importance of highways (if they are used primarily for trucks and buses) — also a “binding” mega project – for linking cities together.

    With regard to 5:
    I think Jacobs’ comments are being highly misinterpreted, and that it’s thus quite untrue that Jacobs opposed rapid transit and supported jitneys as an alternative! Jacobs did point out rapid transit is run as a monopoly – which, by the way, was not “really” true in the early days of rapid transit, at least in NYC – and that this creates problems. She also pointed out that rapid transit planners today don’t seem to know how to plan routes the way the used to (the implication being that rapid transit is good for the health of cities and that planners used to do a better job of it). But that doesn’t mean that she opposed rapid transit – and that she suggested jitneys as an alternative!!! Also, it should be remembered that in her last book she pointed out the economic usefulness of light rail – which is also, likely these days, to be run as a monopoly. Basically speaking, Jacobs felt that the more different modes of transit a place has, the better. The worst thing for a system’s health is to have an overdependence on one form of transportation.
    With regard to 6:
    Although I don’t think Jacobs would see HSR as lessening the independence of cities, she does disagree with you about the benefits of binding ten small cities into one.

    With regard to 7:
    In an interview with the John Tierney, in the New York Times Sunday magazine, both Jacobs and Tierney dispute, so it seems to me, the idea that New York City (and, more importantly the large part of current NYC that is Brooklyn) is better off because of amalgamation.

    Benjamin Hemric
    Thursday, Dec. 1, 2011, 10:05 p.m.

  6. Pingback: Megaregional Change | Pedestrian Observations
  7. Charley Ferrari

    I think you’re right that it’s the same analysis happening on the country/megaregion level as within cities, and IMO the dynamics become clearer when looking at larger regions.

    I’d say this is similar to early vs later analyses on globalization. At first people thought the world would get “flatter”, with people spreading out and location mattering less. As Alon mentioned to the first commenter however, if transportation costs fall (especially with a high capital/low operating cost mode like HSR), and infrastructure is built in response to existing demand, then existing trends will be reinforced.

    This however hinges on the fact that infrastructure is built to reflect current demand. While it’s easier to justify expenditure on existing demand, I can’t help but think that infrastructure is a long term investment, and should be tilted way more towards future demand. Transit and rail was originally built up privately and might not be the best example, but I’d say it’s more comparable to how the NYC street grid was originally placed over farmland anticipating future dense growth. The skeleton came first, and future growth was shaped by it.

    I’m not sure what this would mean for HSR in particular. Instead of planning HSR for the Northeast corrider today, what would change if it were planned for a future version of the corrider subjugated by its larger cities? Interesting to think about.

  8. Pingback: Linear Compression: How HSR is Like Rapid Transit | Pedestrian Observations

Leave a Reply to Stephen Smith Cancel reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.